Development of Smart Pavement Design Sensitivity Analysis Software for Asset Management System

: There are a number of pavement management systems, but most of them are limited in providing pavement design and pavement design sensitivity information. This paper presents e ﬀ orts towards the integrated pavement design and management system, by developing smart pavement design sensitivity analysis software. In this paper, the sensitivity analyses of critical design input parameters have been performed to identify input parameters which have the most signiﬁcant impacts on the pavement thickness. Based on the existing pavement design procedures and their sensitivity analysis results, a smart pavement design sensitivity analysis (PDSA) software package was developed, to allow a user to retrieve the most appropriate pavement thickness and immediately perform pavement design sensitivity analysis. The PDSA software is a useful tool for managing pavements, by allowing a user to instantaneously retrieve a pavement design for a given condition from the database and perform a design sensitivity analysis without running actual pavement design programs. The proposed smart PDSA software would result in the most e ﬃ cient pavement management system, by incorporating the optimum pavement thickness as part of the pavement management process.


Introduction
To minimize the life cycle cost (LCC) of building and maintaining pavements, it is critical to determine the most appropriate pavement thickness for the given traffic level, subgrade condition and environmental factor. It is essential to evaluate the impacts of design on construction and maintenance, in order to obtain minimum pavement life-cycle costs. [1]. Main objective of this research is to provide pavement managers with a tool which performs a pavement thickness design sensitivity analysis on the fly, which would result in the minimum LCC of pavements.
There exist a number of pavement design software packages that generate different thicknesses for the given conditions. Most of these pavement design software packages do not give the user an option to perform the design sensitivity analysis by automatically varying input values. Therefore, the impact of each design parameter on pavement design is not transparent to a user. Therefore, there is a critical need for comparing these existing pavement design software packages and identifying the critical design input parameters through the design sensitivity analysis process.
First, the following two pavement design software packages were evaluated with respect to how they are different in determining design input parameters and their influences on the pavement thicknesses: the StreetPave software, based on the American Concrete Pavement Association (ACPA) thickness design for concrete highway and street pavements, and the Portland Cement Association

Survey of Adjoining Five States about Pavement Design Procedures for Low-Volume Roads
AASHTO 1993 pavement design guide lists a minimum asphalt pavement thickness as 1.0 inch and the minimum concrete pavement thickness as 5.0 inches for the lowest traffic level, ranging from 50,000 to 100,000 equivalent single axle loads (ESALs). However, for the similar traffic level, the asphalt institute recommends a minimum of 3.0 inches for asphalt pavement and the PCA recommends a minimum of 7.0 inches of concrete pavement. ACPA recommends a lower limit for concrete pavement thickness of 4.0 inches for automobiles and 5.0 inches for limited truck traffic. According to the ACPA design table, a minimum concrete pavement thickness for a light residential street is 4.0 inches.
To learn more about pavement design procedures for the low volume roads, five adjoining state departments of transportation (Minnesota, Wisconsin, Illinois, Missouri and South Dakota) were surveyed. As shown in Table 1, a questionnaire was prepared to identify pavement thickness design methods and their common input parameters for local roads. The survey responses from five state Departments of Transportation (DOT) are summarized in Table 2. As can be seen from Table 2, five states differ in pavement design procedures, typical pavement design thicknesses, and reliability factors. However, they agreed on the important design factors as traffic loading and subgrade. These survey results have been used to identify the critical pavement design input parameters, and their ranges of typical values. Table 1. Survey questions with respect to pavement design procedures for low-volume roads.

No. Question
1 What kind of pavement design methodology do you use for local road? 2 What are the layer types and thicknesses of a typical local road (asphalt pavement) in your states? 3 What are the layer types and thicknesses of a typical local road (concrete pavement) in your state? 4 What are the most important factors for designing local road? 5 What type of soil is most common in your state? 6 What type of traffic input do you use for local road design? 7 What level of the design reliability does your state agency use for local road?

Sensitivity Analysis of Design Input Parameters for StreetPave and WinPAS
The following two pavement design software packages were evaluated on how they are different in inputting design input parameters and their influences on the pavement thickness: StreetPave and WinPAS software. The StreetPave is concrete pavement thickness design software developed by ACPA [12]. ACPA later incorporated an asphalt pavement design process based on the Asphalt Institute method, to generate an equivalent asphalt design for the given load carrying capacity requirement. The WinPAS software performs pavement thickness design based on the AASHTO 1993 pavement design guide [13]. As shown in Figure 1, four critical design input parameters were identified, and the sensitivity analysis of each parameter was performed using each of two pavement design software packages. Typical design input parameters were input into both StreetPave and WinPAS software, and the sensitivity analyses of various pavement design parameters have been performed.

Design Input Parameters
The input design parameters that affect the pavement thickness the most were identified as a design life, a truck traffic volume, a reliability and a soil support. A design life is selected as either 20 or 40 years and a reliability is selected as 50%, 80%, or 90%, which are the typical minimum and maximum ranges for local streets. For a reliability, additional value of 80% was added to determine if the impacts of reliability values would be nonlinear. The WinPAS, which is based on the AASHTO 1993 pavement design guide, accepts a reliability directly into the design equation. A reliability is adopted for StreetPave by applying a reliability to the flexural fatigue equation for concrete pavements, and to the resilient modulus of the subgrade for asphalt pavements. A truck traffic volume was selected to represent various traffic loading levels, which were converted to the equivalent single axle load (ESAL) as a design input. Typical truck traffic volume is estimated for each functional classification of roads by multiplying typical average daily traffic (ADT) and a typical percentage of trucks. Table 3 shows the typical average daily truck traffic (ADTT) values as a traffic input parameter.

Design Input Parameters
The input design parameters that affect the pavement thickness the most were identified as a design life, a truck traffic volume, a reliability and a soil support. A design life is selected as either 20 or 40 years and a reliability is selected as 50%, 80%, or 90%, which are the typical minimum and maximum ranges for local streets. For a reliability, additional value of 80% was added to determine if the impacts of reliability values would be nonlinear. The WinPAS, which is based on the AASHTO 1993 pavement design guide, accepts a reliability directly into the design equation. A reliability is adopted for StreetPave by applying a reliability to the flexural fatigue equation for concrete pavements, and to the resilient modulus of the subgrade for asphalt pavements. A truck traffic volume was selected to represent various traffic loading levels, which were converted to the equivalent single axle load (ESAL) as a design input. Typical truck traffic volume is estimated for each functional classification of roads by multiplying typical average daily traffic (ADT) and a typical percentage of trucks. Table 3 shows the typical average daily truck traffic (ADTT) values as a traffic input parameter.

Pavement Design Sensitivity Analysis Results
As summarized in Table 4, WinPAS software provided a slightly wider range of concrete pavement thickness than StreetPave software, whereas StreetPave provided a wider range of asphalt pavement thickness than WinPAS. The impact of subgrade strength on concrete thickness was similar for both StreetPave and WinPAS software. For both StreetPave and WinPAS software, the increased subgrade strength decreased the concrete pavement thickness by less than 1.0 inch and asphalt pavement thickness by 3.0 inches. As shown in Table 4, when the design life was doubled from 20 to 40 years, StreetPave software increased the thickness by 0.25 inch, but WinPAS software increased the thickness by 1.25 inches. Moreover, in SreetPave, a traffic level exhibited a higher impact on asphalt pavement thickness; up to 2.0 inch more than in WinPAS. Table 4. Comparisons of sensitivity analysis results using StreetPave and WinPAS software.

Type of Pavement System Input Parameter StreetPave Software WinPAS Software
Concrete Pavement Given 80% reliability and 20-year design life, using WinPAS and StreetPave, Figure 2a,b show concrete pavement thicknesses under low vs. high traffic volumes, for low and high subgrade strengths, respectively. As shown in Figure 2a, for the low subgrade strength level, both StreetPave and WinPAS recommended nearly identical concrete pavement thickness. However, for the higher subgrade strength, as shown in Figure 2b, the WinPAS recommended a thicker concrete pavement by 0.5 inch. Overall, the impact of subgrade strength on concrete pavement thickness is greater with StreetPave than WinPAS.
As summarized in Table 4, WinPAS software provided a slightly wider range of concrete pavement thickness than StreetPave software, whereas StreetPave provided a wider range of asphalt pavement thickness than WinPAS. The impact of subgrade strength on concrete thickness was similar for both StreetPave and WinPAS software. For both StreetPave and WinPAS software, the increased subgrade strength decreased the concrete pavement thickness by less than 1.0 inch and asphalt pavement thickness by 3.0 inches. As shown in Table 4, when the design life was doubled from 20 to 40 years, StreetPave software increased the thickness by 0.25 inch, but WinPAS software increased the thickness by 1.25 inches. Moreover, in SreetPave, a traffic level exhibited a higher impact on asphalt pavement thickness; up to 2.0 inch more than in WinPAS.  Given the 80% reliability and 20-year design life, using WinPAS and StreetPave, Figure 3a,b show asphalt pavement thicknesses under low vs. high traffic volumes for low and high subgrade strengths, respectively. As can be seen from Figure 3, WinPAS recommended slightly higher asphalt pavement thicknesses than the StreetPave, except for a condition with a low subgrade strength and a high traffic volume. Given the 80% reliability and 20-year design life, using WinPAS and StreetPave, Figure 3a,b show asphalt pavement thicknesses under low vs. high traffic volumes for low and high subgrade strengths, respectively. As can be seen from Figure 3, WinPAS recommended slightly higher asphalt pavement thicknesses than the StreetPave, except for a condition with a low subgrade strength and a high traffic volume. Given the 80% reliability and 20-year design life, using WinPAS and StreetPave, Figure 3a,b show asphalt pavement thicknesses under low vs. high traffic volumes for low and high subgrade strengths, respectively. As can be seen from Figure 3, WinPAS recommended slightly higher asphalt pavement thicknesses than the StreetPave, except for a condition with a low subgrade strength and a high traffic volume.

Pavement Design Sensitivity Analysis Software
The proposed pavement design sensitivity analysis (PDSA) software is a computer program that combines pavement design function and sensitivity analysis capability. The PDSA software consists of four modules: input pavement design data, asphalt pavement design, concrete pavement design, and design sensitivity analysis. The design data input module provides a spreadsheet user interface to input design data. The asphalt and concrete pavement design modules provide the user with a tool to query the design database. The sensitivity analysis module provides a graphic user interface to run a design sensitivity analysis for various input parameters.

PDSA Database Structure
The PDSA software stores the pavement design data in a Microsoft Access format. Figure 4 shows database entities, attributes, and their relationship, that consists of the main tbDesign table with associated tables, which include tbPavementProperty, tbSubbaseCondition, tbDesignLife, tbTraffic and tbReliability. The tbPavementProperty table stores the basic structural properties of a pavement, including a tbLayer subtable, where the pavement layer data associated with the pavement structure are stored. Similarly, the tbSubbaseCondition, tbTraffic, tbDesignLife, and tbReliability tables are designed to store information about subbase, traffic, design life, and reliability factors, respectively. With an integrated database, the stored pavement design data can easily be retrieved for various design parameters, such as subgrade strength and traffic level.

Pavement Design Sensitivity Analysis Software
The proposed pavement design sensitivity analysis (PDSA) software is a computer program that combines pavement design function and sensitivity analysis capability. The PDSA software consists of four modules: input pavement design data, asphalt pavement design, concrete pavement design, and design sensitivity analysis. The design data input module provides a spreadsheet user interface to input design data. The asphalt and concrete pavement design modules provide the user with a tool to query the design database. The sensitivity analysis module provides a graphic user interface to run a design sensitivity analysis for various input parameters.

PDSA Database Structure
The PDSA software stores the pavement design data in a Microsoft Access format. Figure 4 shows database entities, attributes, and their relationship, that consists of the main tbDesign table with associated tables, which include tbPavementProperty, tbSubbaseCondition, tbDesignLife, tbTraffic Infrastructures 2020, 5, 56 7 of 13 and tbReliability. The tbPavementProperty table stores the basic structural properties of a pavement, including a tbLayer subtable, where the pavement layer data associated with the pavement structure are stored. Similarly, the tbSubbaseCondition, tbTraffic, tbDesignLife, and tbReliability tables are designed to store information about subbase, traffic, design life, and reliability factors, respectively. With an integrated database, the stored pavement design data can easily be retrieved for various design parameters, such as subgrade strength and traffic level.

PDSA Software Flowchart
With an integrated database, the entered data can be reused for various pavement design conditions. Thus, a user is not required to input the same data repeatedly in contrast to other pavement design software packages, most of which do not utilize a database. The pavement design results generated from various pavement design software packages can be saved in the separate databases, and utilized later for pavement design and sensitivity analysis . Figure 5 shows a flow chart for inputting the pavement design data and running a sensitivity analysis. First, the pavement design data obtained by running various pavement design software packages should be entered into the pavement design database. Once the pavement design data is saved in the database, the user can then select design input parameters, such as pavement type, design life, reliability, traffic, and subgrade conditions. The PDSA software runs the structured query language (SQL) query based on the user selections, and then displays sensitivity analysis results for the user to review the sensitivity of the input parameters.

PDSA Software Flowchart
With an integrated database, the entered data can be reused for various pavement design conditions. Thus, a user is not required to input the same data repeatedly in contrast to other pavement design software packages, most of which do not utilize a database. The pavement design results generated from various pavement design software packages can be saved in the separate databases, and utilized later for pavement design and sensitivity analysis. Figure 5 shows a flow chart for inputting the pavement design data and running a sensitivity analysis. First, the pavement design data obtained by running various pavement design software packages should be entered into the pavement design database. Once the pavement design data is saved in the database, the user can then select design input parameters, such as pavement type, design life, reliability, traffic, and subgrade conditions. The PDSA software runs the structured query language (SQL) query based on the user selections, and then displays sensitivity analysis results for the user to review the sensitivity of the input parameters.

PDSA Software Flowchart
With an integrated database, the entered data can be reused for various pavement design conditions. Thus, a user is not required to input the same data repeatedly in contrast to other pavement design software packages, most of which do not utilize a database. The pavement design results generated from various pavement design software packages can be saved in the separate databases, and utilized later for pavement design and sensitivity analysis . Figure 5 shows a flow chart for inputting the pavement design data and running a sensitivity analysis. First, the pavement design data obtained by running various pavement design software packages should be entered into the pavement design database. Once the pavement design data is saved in the database, the user can then select design input parameters, such as pavement type, design life, reliability, traffic, and subgrade conditions. The PDSA software runs the structured query language (SQL) query based on the user selections, and then displays sensitivity analysis results for the user to review the sensitivity of the input parameters.

Pavement Design Data Input Module
The pavement design result generated from other pavement design software packages can be stored in the PDSA database and retrieved later for pavement design and sensitivity analysis. The pavement design data obtained by running various pavement design software packages can be entered into the PDSA pavement design database from the pavement design data input module. As shown in Figure 6, the input screen is designed for a user to enter pavement design data into the database. Once the pavement design data is saved in the database, the user can then retrieve the most appropriate design, given design input parameters such as pavement type, design life, reliability, traffic, and subgrade conditions.

Pavement Design Data Input Module
The pavement design result generated from other pavement design software packages can be stored in the PDSA database and retrieved later for pavement design and sensitivity analysis. The pavement design data obtained by running various pavement design software packages can be entered into the PDSA pavement design database from the pavement design data input module. As shown in Figure 6, the input screen is designed for a user to enter pavement design data into the database. Once the pavement design data is saved in the database, the user can then retrieve the most appropriate design, given design input parameters such as pavement type, design life, reliability, traffic, and subgrade conditions.

Asphalt and Concrete Pavement Thickness Design Modules
After entering all the design data into the database, when a user enters the design input parameters such as traffic, subgrade strength, etc., the PDSA software retrieves the most appropriate pavement thickness from the database for the given condition. As shown in Figure 7, both asphalt and concrete pavement thickness design modules include input screens, to allow a user to input project information and design parameters such as design life, reliability, traffic, and subgrade conditions. Once a user provides necessary input values and chooses the pavement design method, the pavement thickness is displayed by retrieving the most appropriate pavement design value from the database. The project information, design inputs and the design output can then be printed in a report format.

Asphalt and Concrete Pavement Thickness Design Modules
After entering all the design data into the database, when a user enters the design input parameters such as traffic, subgrade strength, etc., the PDSA software retrieves the most appropriate pavement thickness from the database for the given condition. As shown in Figure 7, both asphalt and concrete pavement thickness design modules include input screens, to allow a user to input project information and design parameters such as design life, reliability, traffic, and subgrade conditions. Once a user provides necessary input values and chooses the pavement design method, the pavement thickness is displayed by retrieving the most appropriate pavement design value from the database. The project information, design inputs and the design output can then be printed in a report format.

Pavement Design Input for Sensitivity Analysis Module
Since the pavement design data are already stored in the database, the user can perform the sensitivity analysis instantaneously against pavement design parameters such as design life, reliability, traffic level and subgrade condition. The PDSA software produces a plot of multiple design thicknesses for various input parameters that would allow a user to consider different thicknesses for varying input values. The PDSA software works as a design database and provides the user with an easy-to-use design and sensitivity analysis tool. Figures 8-11 show a sensitivity analysis result, to illustrate the influence of the design life, reliability, traffic, and subgrade support on the asphalt pavement thickness, respectively. Default

Pavement Design Input for Sensitivity Analysis Module
Since the pavement design data are already stored in the database, the user can perform the sensitivity analysis instantaneously against pavement design parameters such as design life, reliability, traffic level and subgrade condition. The PDSA software produces a plot of multiple design thicknesses for various input parameters that would allow a user to consider different thicknesses for varying input values. The PDSA software works as a design database and provides the user with an easy-to-use design and sensitivity analysis tool. Figures 8-11 show a sensitivity analysis result, to illustrate the influence of the design life, reliability, traffic, and subgrade support on the asphalt pavement thickness, respectively. Default values for the design parameters are a design life of 20 years, an AATT of 100, a reliability of 80%, and a soil support value of 2900 psi. For all sensitivity analyses, granular subbase thicknesses are assumed as 6 or 12 inches. By clicking on the tab on the top of the main screen, the sensitivity analysis result based on other input parameters such as design life, traffic level and soil condition can be instantaneously displayed. As shown in Figure 8, for a 20-year design period, the asphalt pavement thicknesses are 8.3 and 8.62 inches when a granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the design life is increased from 20 to 40 years, both asphalt pavement thicknesses are 9.64 inches. Based on this sensitivity analysis, for a small increase in asphalt pavement thickness by 1.34 or 1.02 inches, the design life can be doubled from 20 to 40 years. It can be concluded that, for a 40year design, there is no need to increase a subbase thickness from 6 to 12 inches.
As shown in Figure 9, for 50% reliability, the asphalt pavement thicknesses are 7.31 and 7.6 inches when granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the reliability is increased from 50% to 90%, asphalt pavement thicknesses are 8.98 and 9.34 inches for 12-and 6-inch subbase thicknesses, respectively. Based on this sensitivity analysis, the effect of reliability is quite significant, by increasing the thickness by 1.7 inches when a reliability was increased from 50% to 90%.
As shown in Figure 10, for a very low truck traffic volume, the asphalt pavement thicknesses are 5.26 and 5.34 inches when the granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the ADTT is increased to 800, asphalt pavement thicknesses are more than doubled to 11.77 and 12.11 inches, for 12-and 6-inch subbase thicknesses, respectively. When the ADTT is increased up to 6200, the asphalt pavement thicknesses gradually increased. Based on this sensitivity analysis, the effect of truck traffic volume is very significant, increasing the thickness by up to 11 inches.
As shown in Figure 11, for a very week soil resilient modulus of 1450 psi, the asphalt pavement thicknesses are 9.84 and 10.32 inches when the granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the soil strength is tripled to 4350 psi, asphalt pavement thicknesses are decreased to 7.3 and 7.6 inches for 12 and 6-inch subbase thicknesses, respectively. Based on this sensitivity analysis, the effect of soil strength is significant, decreasing the thickness by up to 2.7 inches.           As shown in Figure 8, for a 20-year design period, the asphalt pavement thicknesses are 8.3 and 8.62 inches when a granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the design life is increased from 20 to 40 years, both asphalt pavement thicknesses are 9.64 inches. Based on this sensitivity analysis, for a small increase in asphalt pavement thickness by 1.34 or 1.02 inches, the design life can be doubled from 20 to 40 years. It can be concluded that, for a 40-year design, there is no need to increase a subbase thickness from 6 to 12 inches.
As shown in Figure 9, for 50% reliability, the asphalt pavement thicknesses are 7.31 and 7.6 inches when granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the reliability is increased from 50% to 90%, asphalt pavement thicknesses are 8.98 and 9.34 inches for 12-and 6-inch subbase thicknesses, respectively. Based on this sensitivity analysis, the effect of reliability is quite significant, by increasing the thickness by 1.7 inches when a reliability was increased from 50% to 90%.
As shown in Figure 10, for a very low truck traffic volume, the asphalt pavement thicknesses are 5.26 and 5.34 inches when the granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the ADTT is increased to 800, asphalt pavement thicknesses are more than doubled to 11.77 and 12.11 inches, for 12-and 6-inch subbase thicknesses, respectively. When the ADTT is increased up to 6200, the asphalt pavement thicknesses gradually increased. Based on this sensitivity analysis, the effect of truck traffic volume is very significant, increasing the thickness by up to 11 inches.
As shown in Figure 11, for a very week soil resilient modulus of 1450 psi, the asphalt pavement thicknesses are 9.84 and 10.32 inches when the granular subbase thicknesses are 12 inches and 6 inches, respectively. However, when the soil strength is tripled to 4350 psi, asphalt pavement thicknesses are decreased to 7.3 and 7.6 inches for 12 and 6-inch subbase thicknesses, respectively. Based on this sensitivity analysis, the effect of soil strength is significant, decreasing the thickness by up to 2.7 inches.
The above four design sensitivity scenarios are for illustration purposes only, and they may not represent all field conditions. For example, as shown in Figure 11, the asphalt pavement thickness decreases linearly as the subgrade support value increases. It should be noted that this linear relationship is valid only for given traffic, reliability and design life. Among four design parameters, it can be concluded that the truck traffic volume has the highest impact on the asphalt pavement thickness.

Summary and Conclusions
This paper discusses how two existing pavement design software packages are different in accepting various design input parameters and their influences on the pavement thickness. StreetPave software designs the concrete pavement thickness based on the PCA method and the equivalent asphalt pavement thickness. The WinPAS software performs concrete and asphalt pavement designs, following the AASHTO 1993 design method. Both software packages were run for various input parameters, such as traffic, subgrade strength, reliability, and design life. The sensitivity analyses of various design input parameters have been performed by using these pavement design software packages to identify which input parameters are most influential on the pavement thickness. Overall, the traffic load has the highest impact on pavement thicknesses, followed by the subgrade strength, reliability and design life.
The design results from StreetPave and WinPAS software were compared with regards to how they are different in determining design input parameter values and their influences on the pavement thickness. For the same input parameters, StreetPave software recommended a thicker asphalt pavement than WinPAS. For the high subgrade strength and low traffic volume, however, WinPAS software recommended a thicker asphalt pavement than the StreetPave. A sensitivity analysis revealed that these pavement design software packages may recommend slightly different pavement thicknesses for a similar condition.
There are numerous pavement design software packages available, but none of them have the capability of storing the pavement design thicknesses for the future use of similar conditions or perform the design sensitivity analysis. Most pavement design software packages allow a user to design one pavement section at a time, and some software takes extensive time to run. To expedite the pavement design and sensitivity analysis, a smart pavement design sensitivity analysis (PDSA) software was developed.
The PDSA is a database system, where a user can retrieve previously run design data using various pavement design software packages. Since the pavement input data can be categorized in finite number, say up to five categories, there is a finite set of input data sets. Therefore, all input data sets can be run beforehand, and the pavement design results can be stored in the database for a quick